From: shunsuke (pstade.mb_at_[hidden])
Date: 2007-11-02 07:22:09
Corrado Zoccolo wrote:
> I don't think we can have a generic nonnullable<Z> that can select and
> forward only the correct constructors. You have to specialize it for every
> pointer-like class, so it doesn't simplify the developer work,
Right, it would lead to complexity in order to
know the result type of dereference.
So, why not nonnullable_poly(or something).
boost::any and boost::function are nullable but
those names don't have "optional_" prefix.
Regards,
-- Shunsuke Sogame